SAN DIEGO-Weighing whether or not to perform tonsillectomy boils down to a balance between benefit and harm, declared Richard Rosenfeld, MD, MPH, at the lively and well-attended miniseminar on evidence-based tonsillectomy at the American Academy of Otolaryngology-Head and Neck Surgery annual meeting here.
The session was designed to help guide otolaryngologists in their decision-making. Even though the surgery is still commonly performed, Dr. Rosenfeld reminded his audience, it remains a nontrivial one.
Dr. Rosenfeld, Chairman and Professor of Otolaryngology at SUNY Downstate Medical Center in Brooklyn, NY, and his co-presenter, Martin J. Burton, MA, DM, of the Department of Otolaryngology- Head and Neck Surgery at Oxford Radcliffe Hospitals in the UK and founding editor of the Cochrane Collaboration ENT Section, summarized the dilemmas surrounding tonsillectomy, including appropriate indications, the most effective surgical methods, perioperative management, and clinical outcomes. Both presenters used humor to good effect during their well-organized presentation designed to introduce a more rational approach to tonsillectomy.
Tonsillectomy is not as common today as it was during the mid-20th century, when 1.5 million tonsillectomies were performed each year in the United States (that number had decreased to approximately 500,000 by 2000). Despite the abundance of research studies-770 randomized controlled trials, more than 440 of which have been published since 2000-there is still uncertainty regarding many aspects of tonsillectomy, including patient selection. Otolaryngologists are familiar with the Pittsburgh criteria advanced by Paradise et al. in the mid 1980s1 (7 throat infections in the past year, 5 per year for two years, or 3 per year for three years) and many use these criteria or a more relaxed version to guide their surgical decision-making.
However, other human and psychological factors can affect surgeons’ decisions. Mr. Burton cited results from a 2000 UK survey questionnaire, in which otolaryngologic surgeons were queried about factors that most influence their decisions to perform tonsillectomy. Surprisingly frank, the respondents reported that 50% of the time, assertive parents determined their decision to opt for a tonsillectomy.
Rarely Black or White
After a brief history of the indications for tonsillectomy-an operation that has existed since ancient times-Mr. Burton offered a panoply of the review efforts designed to ferret out the most appropriate use of tonsillectomy, as well as the best surgical techniques and perioperative management strategies. The Cochrane Collaboration, initiated by Sir Iain Chalmers, conducts systematic reviews of studies to help facilitate rational health care decision making. After comprehensive searches of the medical literature on a designated topic (using databases such as The Cochrane Library, MEDLINE, and EMBASE), editors cull through groups of relevant studies and rate the level of evidence (with randomized controlled trials [RCTs] ranked as level 1 evidence). Findings from studies that meet inclusion criteria (ie, that all address the same research question) are then combined and analyzed in the aggregate in order to arrive at general conclusions about the effects of interventions.
Mr. Burton was lead author, along with another colleague, of the most recently published systematic review of tonsillectomy or adenotonsillectomy versus nonsurgical treatment for chronic-recurrent acute tonsillitis.2 A compendium of five RCTs with a total of 719 children and 70 adult participants, the review did reveal a modest advantage for tonsillectomy over nonsurgical treatment-and this was for the children with the most serious disease according to the Pittsburgh criteria. As is often observed in clinical practice, however, many children with more moderate infections simply got better over time without surgery. This review, published in January, was an updated version of an earlier review examining the same question, and was slightly more definitive. Even though the 2009 review revealed a modest effect for removing the tonsils in the subgroup of severely affected children, its conclusions do not tell the full story, said Mr. Burton. The predictability of postoperative sore throat episodes may actually be a gain for parents and families, he said, and might be preferred over the disruptive effects of the unpredictable illness episodes over the span of a year.
Even systematic reviews do not offer answers that are clearly black or white, said Mr. Burton. Earlier, he pointed out another drawback of relying on RCTs and systematic reviews for definitive answers regarding tonsillectomy. Sample sizes are often too small to detect rare outcomes, he said. (This was the case for a Cochrane review comparing surgical techniques.3) To tease out rare outcomes, one can turn to other study designs, which, although not randomized, are a perfectly legitimate vehicle for deciphering this information. One such example was the National Prospective Tonsillectomy Audit, conducted in the UK between July 2003 and July 2004, which included more than 30,000 patients. The data, said Mr. Burton, are the best we will ever get to answer the question, ‘What are the risks of rare events after tonsillectomy?’ The results from the audit revealed that, compared to cold steel dissection, a variety of different surgical techniques (including monothermy, diathermy, and coblation) all carried much higher risks of postoperative hemorrhage. The risk of hemorrhage after coblation, for example, was three times higher than after cold steel dissection.
When presenting these data, Mr. Burton invited his audience to consider the tradeoffs implicit in their surgical decisions. If you could show that your new technique or new wonderful instrument was better in terms of less pain, how much better does it have to be to warrant an increase in hemorrhage? I don’t know the answer to that, he admitted, but you need to ask that question for yourselves.
Mr. Burton summarized results from reviews of perioperative management studies, which also revealed equivocal results. For instance, use of nonsteroidal anti-inflammatory drugs (NSAIDs) did not appear to increase bleeding, but the number of participants was small (13 trials with 155 children), so it is possible that rare bleeding events were not detected. In another review of nine trials, antibiotics significantly reduced fever, but had no effect on pain reduction. He also cited the dexamethasone study,4 undertaken to study the drug’s efficacy in decreasing post operative nausea and vomiting in children undergoing tonsillectomy. That study was halted early because although postoperative nausea and vomiting were reduced, there was a disturbing increase in postoperative hemorrhage in the group randomized to dexamethasone.
The Pittsburgh Criteria Backstory
Dr. Rosenfeld then took the floor to address the integration of evidence with individual patient care. He noted that before Paradise published the 1984 study showing a benefit for tonsillectomy, the researcher had conducted an earlier study following children to monitor their episodes of infection. Recruited via local media and referrals from pediatricians, the children met the Pittsburgh criteria and were observed for the next year. Sure enough, said Dr. Rosenfeld, almost half the kids had one or fewer episodes [of throat infection] in the next year. So, the natural history is good.
There is no doubt, he conceded, that the sicker the child, the more impact the surgery will have. It remains difficult, however, to show real benefit after surgery for those who are only mildly symptomatic. On the other hand, the risks of tonsillectomy are real, said Dr. Rosenfeld. In addition to the more common side effects of pain and postoperative hemorrhage, there are uncommon risks, such as endotracheal tube ignitions, cervical facial emphysema, bacterial meningitis, and psychological trauma. Surgeons would do well, he said, to always consider the magnitude of the expected benefit against the risks.
The decrease in tonsillectomy procedures has, however, introduced another iatrogenic consequence, according to Dr. Rosenfeld: sleep-disordered breathing. This is a disorder that was barely mentioned 50 years ago, but now represents the primary indication for tonsillectomy. The current trials weighing surgical versus nonsurgical treatment for sleep apnea, he said, fail to address quality of life and neurocognitive sequelae (such as failure to thrive, nocturnal enuresis, underperformance at school).
The decisions for surgeons remain complex, Dr. Rosenfeld concluded. If you were hoping for an easy decision, an easy strategy, we’ve probably not provided that here. He urged participants to include their expertise and patient preference in the decision-making equation. He also cautioned his colleagues to resist gizmo idolatry as described in a recent paper5-that the newest technology is always superior-and to thoughtfully consider the evidence, indications and complications that go along with this nontrivial surgery.
There are valid reasons for this surgery, said Dr. Rosenfeld, who concluded, I always tell my residents that the best tonsillectomy is one that is prudently avoided.
- Paradise JL, Bluestone CD, Bachman RZ, et al. Efficacy of tonsillectomy for recurrent throat infection in severely affected children. Results of parallel randomized and nonrandomized clinical trials. N Engl J Med 1984;310:674-83.
- Burton MJ, Glasziou PP. Tonsillectomy or adeno-tonsillectomy versus non-surgical treatment for chronic/recurrent acute tonsillitis. Cochrane Database Syst Rev 2009;1:DC001802.
- Burton MJ, Doree C. Coblation versus other surgical techniques for tonsillectomy. Cochrane Database Syst Rev 2007;3:CD004619.
- Czarnetzki E, Elia N, Lysakowski C, et al. Dexamethasone and risk of nausea and vomiting and postoperative bleeding after tonsillectomy in children: a randomized trial. JAMA 2008;300:2621-30.
- Leff B, Finucane TE. Gizmo idolatry. JAMA 2008;299:1830-2.
©2009 The Triological Society